Comparison of the effects of litter decomposition process on soil erosion under simulated rainfall

Study area description

Yangtze River Basin is situated in central China (Fig. 1). Its geographical coordinates are between 30° 48′ 30″–31° 02′ 30″ N and 112° 48′ 45″–113° 03′ 45″ E. Taizishan is located in the transition zone between the north and south of China, with an altitude of 403–467.4 m. It belongs to the subtropical monsoon humid climate zone and has obvious karst landforms. The farm area is 7576 hectares, the forest coverage rate is 82.0%, and the vegetation is mainly Masson pine, fir, and various broad-leaved tree species. Increased forest coverage reduces sediment production³⁰. The soil is mainly viscous yellow–brown soil and loess parent material. Rain is concentrated in summer, with an average annual rainfall of 1094.6 mm and an average annual temperature of 16.4 °C. Rainfall-related flood risk increased in the Yangtze River Delta in recent years³¹.The study was based in a Pinus massoniana forest in the Taizishan forest farm of Hubei Province. The Pinus massoniana (Masson pine) is a common species distributed in Central China.

Experiment design

We chose the Pinus massoniana forest with 47a in the study area as the research object. In the typical Pinus massoniana forest, the separate layers of litter (semi-decomposed and non-decomposed layers) were collected from several 1 m × 1 m quadrat and placed in grid bags. The litter of the semi-decomposed layer have no complete outline, and the color was brown. As the litter leaves of the completely decomposed layer are powdery and are combined with the soil layer, this layer is difficult to collect. Before testing, it was necessary to clean the soil off the pine needles and then allow the litter to dry naturally. The characteristics of the semi-decomposed and non-decomposed litter layers are shown in Table 1. The soil samples need to be dried and screened by 10 mm. When filling the soil trough, every 0.1 m of soil thickness was one layer, for a total of four layers (0.4 m). The characteristics by soil particle sizes are different (Fig. 2). The soil samples were dried naturally, crushed, and then sieved. The soil trough (2 m long, 0.5 m wide and 0.5 m deep) was filled to have a bulk density of 1.53 g·m⁻³. In this process, an appropriate amount of water was sprinkled on the surface of each soil layer to achieve a soil moisture content consistent with the surrounding, undisturbed, or natural, state. The simulation experiment was conducted in the Jiufeng rainfall laboratory at Beijing Forestry University, China. We used a rainfall simulation system (QYJY-503T, Qingyuan Measurement Technology, Xi’an, China) used a rotary downward spray nozzle. The system is able to simulate a wide range of rainfall intensities (10 to 300 mm h⁻¹) using various water pressure and nozzle sizes controlled by a computer system.

According to the results of the field forest investigation, the litter was covered with the experimental treatments shown in Table 2. The treatments mass coverage of non-decomposed litter layer was named as follows: N1 denoted litter mass coverage 0 g·m⁻², N2 was ‘the non-decomposed litter mass coverage 100 g·m⁻²’, N3 was ‘the non-decomposed litter mass coverage 200 g·m⁻²’, and N4 was ‘the non-decomposed litter mass coverage 400 g·m⁻²’, N5 was ‘the semi-decomposed litter mass coverage 100 g·m⁻²’, N6 was ‘the non-decomposed litter mass coverage 100 g·m⁻² and the semi-decomposed litter mass coverage 100 g·m⁻²’, N7 was ‘the non-decomposed litter mass coverage 200 g·m⁻² and the semi-decomposed litter mass coverage 100 g·m⁻²’. N2, N3 and N4 were the undissolved state of litter layer, and N4 (non-decomposed state, ND), N7 (initial stage of litter decomposition, ID), N6 (middle stage of litter decomposition, MD) and N5 (final stage of litter decomposition, FD) respectively represent different stages of litter decomposition.

According to the rainfall in the Taizishan area of Hubei Province, erosive rainfall and extreme rainstorms were selected as the research conditions. Summer rainfall events occur mainly in the summer in this area, and a rainfall intensity of 60 mm·h⁻¹ was the most common erosive rainfall intensity. Under extreme weather conditions, the rainfall intensity can reach up to 120 mm·h⁻¹. Our experiments were conducted with 60 and 120 mm·h⁻¹ rain intensities with a rainfall that lasted 1 h. According to the field investigation data of forest land, this area is a low mountain and hilly area with a slope mostly between 5° and 10°. Therefore, 5° and 10° were selected for the slope treatments in this study. The combination of slope and rainfall intensity was named as follows: T1 denoted ‘Slope 5° and rainfall intensity 60 mm·h⁻¹’, T2 was ‘Slope 10° and rainfall intensity 60 mm·h⁻¹’, T3 was ‘Slope 5° and rainfall intensity 120 mm·h⁻¹’, and T4 was ‘Slope 10° and rainfall intensity 120 mm·h⁻¹’. With two rainfall intensities, two slopes, seven litter coverage gradient and two repetitions combined, this study had a total of 56 rainfall events.

Experimental procedure

Before the test, the soil samples were wetted for 10 h and then drained for 2 h to eliminate the effect of the initial soil moisture on the soil detachment measurement. When the simulated rainfall started, all the runoff and sediment produced from plot were collected every 5 min in the first 10 min, and then collected once every 10 min during the subsequent 50 min. At the same time, runoff velocity, depth and temperature were measured and vernier calliper (accuracy 0.02 mm) respectively.

The overland flow velocity was measured using dying method (KMnO₄ solution)³². After judging the flow pattern, we confirmed the correction coefficient K value (in laminar flow state, K = 0.67; transition flow state, K = 0.70; turbulent flow state, K = 0.8). The average velocity of overland flow was obtained by multiplying the correction coefficient K and the instantaneous velocity. Runoff depth was measured using vernier calliper (accuracy 0.02 mm). Runoff temperature was measured using thermometer. When the rainfall experiment finished, the collected runoff samples were measured volumetric cylinder and then settled for at least 12 h. The clear water was decanted, and the samples were put into an oven to dry for 24 h under 105 °C. The sediment sample was dried and weighed with an electronic scale.

Calculation of hydrodynamic parameters

Overland flow has the characteristics of a thin water layer, large fluctuations of the underlying surface, and unstable flow velocity. At present, most scholars use open-channel flow theory to study overland flow^33,34. In open-channel flow theory, the Reynold’s number (Re), Froude constant (Fr), flow index (m), resistance coefficient (f), and soil separation rate (({D}_{r})) are the basic parameters of overland flow dynamics, through Reynold’s number (Re), Froude constant (Fr), flow index (m) can distinguish flow patterns. Re is calculated as:

where Re is the Reynolds number of the water flow, which is dimensionless, and can be used to judge the flow state of overland flow. When Re ≤ 500, the flow pattern is laminar; when 500  <  Re ≤ 5000, the flow pattern is transitional; when Re > 5000, the flow pattern is turbulent. R is the hydraulic radius (m), which is generally replaced by flow depth as measured by a vernier calliper (accuracy 0.02 mm). (V) is the average velocity (m·s⁻¹); (nu) is the kinematic viscosity coefficient (m²·s⁻¹), and the calculation formula is (nu) = 0.01775·10⁻⁴·(1 + 0.0337 t + 0.00021 t²), where t is the test overland flow temperature³⁵.

Fr is the Froude constant, which is the ratio of the inertial force to gravity and can be used to distinguish overland flow as rapid flow, slow flow, or critical flow. When Fr < 1, the fluid is considered as slow flow; when Fr = 1, the fluid is critical flow; when Fr > 1, the fluid is rapid flow.

Fr is calculated as:

where (Fr) is the Froude constant of the water flow, which is dimensionless; (V) is the average velocity (m·s⁻¹); g is the acceleration of gravity and has a constant value of 9.8 m·s⁻²; R is a hydraulic radius (m), and is generally replaced by flow depth as measured by a vernier calliper (accuracy 0.02 mm).

Regression fitting is made for runoff depth (h) and single width flow (Q). The runoff depth equation for slope is as follows:

where q is the single width flow (L·m⁻¹·s⁻¹); h is the depth of water on the slope (m); and m is the flow index, which reflects the turbulent characteristics of the flow state. The larger m is, the more energy the flow consumes in the work of resistance. The comprehensive index (k) reflects the characteristics of the underlying surface and the water viscosity of the slope flow. The larger k is, the stronger the surface material of the slope works on the flow.

The resistance of overland flow reflects the inhibition effect of different underlying surface conditions on the velocity of overland flow. The Darcy–Weisbach formula is widely used in research because of its two advantages: applicability and dimensionlessness under laminar and turbulent flow conditions^36,37.

The resistance coefficient (f) is calculated as follows:

where the resistance coefficient f has no dimension; g is the acceleration of gravity and is always 9.8 m·s⁻²; R is a hydraulic radius (m), generally replaced by flow depth measured by a vernier calliper (accuracy 0.02 mm); (V) is the average velocity (m·s⁻¹); and J is the hydraulic gradient, which can be converted by the gradient in a uniform flow state and is generally replaced by the sine value of the gradient.

Shear stress ((tau)) is the main driving force that affects the stripping of soil particles from the surface soil³⁸. Shear stress is calculated as:

where (tau) is the shear force of runoff (Pa); and r is the density of water and sediment concentration flow (kg·m⁻³). This study used a muddy water mass and volume ratio in the unseparated state to calculate the density of water and sediment concentration flow.

Flow power (W) is the runoff power per unit area of water and refers to the power consumed by the weight of water acting on the riverbed surface to transport runoff and sediment. W is calculated as:

where W is the flow power (N·m⁻¹·s⁻¹); and (tau) is the shear force of runoff (Pa).

Soil separation rate (({D}_{r})) refers to the quality of soil in which soil particles are separated from the soil per unit time. The calculation formula is as follows:

where ({D}_{r}) is the rate of soil separation (kg·m⁻²·s⁻¹); ({W}_{w}) is the dry weight of soil before the test; ({W}_{d}) is the dry weight of soil after the test, measured by the drying method (kg); t is the scouring time (s); and A is the surface area of the soil sample (m²).

Source: Ecology - nature.com